Swamy Gadila

Key Skills

Software

Mercor

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Task Types

Entity Ner Classification

Question Answering

Evaluation Rating

Freelancer Overview

I specialize in building and maintaining AI systems with a strong focus on data quality, reliability, and efficient model deployment. My experience spans NLP, computer vision, and retrieval-augmented generation (RAG) projects, where I have worked extensively on developing and optimizing data pipelines, feature engineering, and monitoring data drift in production environments. I am skilled in tools such as PyTorch, Hugging Face, PySpark, SQL, and various MLOps frameworks, and have hands-on experience with vector databases, model fine-tuning, and data preprocessing for both structured and unstructured datasets. My work often involves ensuring the integrity and accuracy of training data, implementing quality assessment pipelines for image data, and designing systems for robust model evaluation and rollback. I am passionate about creating AI solutions that are not only performant but also dependable, with a strong emphasis on data annotation, monitoring, and continuous improvement.

Labeling Experience

Mercor – Medical NLP & Clinical Text Evaluation

MercorTextEntity Ner ClassificationQuestion Answering

I supported a medical NLP project with Mercor by evaluating model outputs on clinical datasets. I labeled medical entities, annotated text related to diagnoses, and classified clinical summaries to ensure they were accurate, relevant, and safe. I reviewed the answers generated by the model to check for correctness and compliance with medical guidelines. My focus was on finding hallucinations, missing clinical details, and unclear phrasing. I provided precise annotations that helped improve model performance and lowered error rates in the evaluation sets.

I supported a medical NLP project with Mercor by evaluating model outputs on clinical datasets. I labeled medical entities, annotated text related to diagnoses, and classified clinical summaries to ensure they were accurate, relevant, and safe. I reviewed the answers generated by the model to check for correctness and compliance with medical guidelines. My focus was on finding hallucinations, missing clinical details, and unclear phrasing. I provided precise annotations that helped improve model performance and lowered error rates in the evaluation sets.

2025 - 2025

Education

Work History